In commercial and transactional printers, it is common to estimate ink usage to determine one of the major components of the cost to print a job. Conventional ink estimation methods involve having to first perform a rasterization (or RIP) of the print job to produce a contone image, which is subsequently halftoned with the same halftone producing algorithms and settings to be employed by the targeted printer. The result is a bitmap from the halftoning operation that describes the resulting drop size for each pixel. The bitmap encodes the different drop sizes using a unique symbol for each different drop size (e.g., level zero for no drop, one for small, two for medium and three for large).
In an actual printer this bitmap data would be the input to the drivers of ink jet printheads. Hence the data used in an actual printer is the same as the data used to estimate a print job. Since the drop sizes for an ink jet are known, the amount of ink required to print the job for each color may be calculated as the sum of ink for each drop size, page and color.
However, the above-described ink estimation process is computationally intensive. Methods to speed up the process involve estimating the ink for a down sampled image and multiplying the result by the down sampling factor. For example, down sampling the contone image data by a factor of two in both directions results in a bitmap having one quarter of the pels of the original. An estimate using the sum of the ink drops, based on the downsampled image must be scaled by a factor of four to obtain an estimate for the ink usage of the original contone image. Yet this process is inefficient since the down sampling factor can only be made so large (e.g., normally 2 or 4 depending on a RIP resolution) before the bitmap resolution/quality is so low/degraded that the estimate becomes too erroneous.
On modern printers there are a variety of settings in both the Digital Front End (DFE) and in the actual print engine that can be independently controlled or modified. Generally, modifying any one of these settings can impact the resultant rasterization (RIP) and subsequent halftone. Thus, any slight change in a printer configuration would necessitate a recalculation of any estimate that the user may wish to be reflective of the changed settings. This can lead to an excessive amount of time and resources being used for estimation when something as simple as, for example, the print speed or resolution changes.
Accordingly, an improved mechanism to perform ink estimation is desired.